No, I am not going to tell you that should be mining the moon. Your instincts about this, like most of the other topics, are correct.
One of the problems with posing a "bold new plan" is that you can't just extrapolate from previous plans. The initial purpose of NASA was to catch up to the Russians. Then Kennedy upped the ante by proposing that we go to the moon. These days the Russians are our partners (in space, at least), and we've been to the moon. So, what next?
The nature of the previous challenges makes people think that the next grand plan also has to be a destination. Mars is a long way away, and we've been to the moon already. Furthermore, simply going to those places is not all that useful. We can accomplish more science by sending unmanned probes.
So, here is my bold plan for NASA, built around three key themes—Explore, Protect, and Innovate. The first one, Explore, is basically the current NASA unmanned programs, which I think are quite good for science content and value. Missions like NGT—the Next Generation Telescope—and the Keppler satellite that will look for Earth-sized planets around other stars are serious, peer-reviewed science.
Protect is about NEOs—the asteroid threat. You are absolutely correct that this is a worthy mission, although I must chide you for being so dismissive. NEO impacts have caused mass extinctions on Earth multiple times. It is absolutely certain that the Earth will be struck again and that it will cause massive destruction.
Indeed we already have one candidate. There is a one in 300 chance that Earth will be struck on March 16, 2880, by an asteroid large enough to destroy civilization and possibly cause the extinction of the human race. But, on the bright side, Prince could re-release his hit song with the new refrain "We're gonna party like its twenty-eight seventy-nine."
A risk 800 years away seems remote, but the danger is that other asteroids are bearing down us on much sooner. An impact with the same energy as the Hiroshima atomic bomb happens every year, but since they tend to explode in the upper atmosphere, the damage isn't obvious. Larger impacts do hit the ground—in 1908, an asteroid hit Tunguska, a remote area in Siberia, and exploded with an energy estimated at 10 megatons—equivalent to the most powerful nuclear weapon in the U.S. arsenal. It flattened trees across an area of 1,000 square miles.
It happened to hit a remote area, in a country in political turmoil, so it wasn't seriously studied for decades. If the impact had occurred in a populated area, or even an unpopulated area in the American West, we would all have grown up with a deep appreciation for the serious consequences of asteroid impact.
There are many other examples—10,000 people were killed by a meteor impact in China in 1490. Millions of people around the world died from climatic shifts from an impact in 535 A.D.—ushering in the Dark Ages. The risk is real, but just the evidence is just far enough removed from our normal attention span that can comfortably ignore.
Most estimates of the mortality risk posed by asteroid impacts put it at about the same risk as flying in a commercial airliner. However, you have to remember that this is like the entire human race riding the plane—it is one of the few risks that really could wipe us all out. The mortality-rate estimate belies our actions—we spend billions of dollars a year on air safety, yet we spend peanuts on the asteroid threat. That is just stupid.
So, you ask what value space flight is. Well, with absolute certainty I can say that it is the only hope of saving the human race. Sooner or later the space program will need to save us by detecting and deflecting an incoming asteroid.
That said, the Protect mission is not a shuttle replacement in the near term. The first task is to understand how many asteroids are out there. The best way to do this is a large ground based telescope (for example, see this). The point of the Protect mission is to undertake work of this sort, as well as a variety of other research so we know what the hell to do on the day we find the next killer asteroid coming at us.
This gets me to Innovate—which is the most interesting one. I think that NASA should be charged with designing and building a new launch platform with a factor-of-10 lower cost than today's best. This means a cheap manned flight system and a cheap unmanned flight system—which may or may not share components. The 10X lower cost must be based on the best current systems—i.e., not the shuttle.
Note that this is somewhat different than your call for new propulsion research. Yes, we need to do research on radical new propulsion—things that could be 100 times cheaper. But when it comes to "only" a 10X improvement, you are within the range of development rather than speculative research. Or, anyway that is what I am betting on.
Is it achievable? Well, it's at least as plausible as going to the moon was in 1962—no NASA committee ever would have proposed that plan. I think that it is feasible because the key elements of space flight are very old ideas, developed by incredibly cost insensitive clients—the military of the United States and U.S.S.R. Yes, there is some commercial space flight, but virtually all of it is a commercialization of previous military technology. Starting from scratch, with cost as a primary goal, it seems likely you wring out a 10X increase in efficiency. That's certainly true if you look at other technology developed by the government, and I am guessing that it is true here, too.
If it could be achieved, it would certainly be worthwhile. A 10X cost improvement would dramatically change the equation for virtually all uses of space—commercial satellites, exploration of the solar system, and asteroid defense. It is hard to overstate the impact that 10X cheaper launches would have on the space program. Commercial uses of space would be feasible. Research could be done on a much broader scale.
Picking apart this vision, or any other, will go beyond the scope of a Slate "Dialogue." It's certainly been fun discussing this with you, Gregg.